1 # This file is part of NIT ( http://www.nitlanguage.org ).
3 # Licensed under the Apache License, Version 2.0 (the "License");
4 # you may not use this file except in compliance with the License.
5 # You may obtain a copy of the License at
7 # http://www.apache.org/licenses/LICENSE-2.0
9 # Unless required by applicable law or agreed to in writing, software
10 # distributed under the License is distributed on an "AS IS" BASIS,
11 # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
12 # See the License for the specific language governing permissions and
13 # limitations under the License.
15 # Simple numerical statistical analysis and presentation
20 # A counter counts occurrences of things
21 # Use this instead of a `HashMap[E, Int]`
24 # var c = new Counter[String].from(["a", "a", "b", "b", "b", "c"])
31 # The counter class can also be used to gather statistical informations.
34 # assert c.length == 3 # because 3 distinct values
35 # assert c.max == "b" # because "b" has the most count (3)
36 # assert c.avg == 2.0 # because it is the mean of the counts
38 class Counter[E
: Object]
41 # Total number of counted occurrences
44 # var c = new Counter[String]
46 # c.inc_all(["a", "a", "b", "b", "b", "c"])
51 private var map
= new HashMap[E
, Int]
53 redef fun iterator
do return map
.iterator
55 # The number of counted occurrences of `e`
56 redef fun [](e
: E
): Int
59 if map
.has_key
(e
) then return map
[e
]
63 redef fun []=(e
: E
, value
: Int)
70 redef fun keys
do return map
.keys
72 redef fun values
do return map
.values
74 redef fun length
do return map
.length
76 redef fun is_empty
do return map
.is_empty
83 # Count one more occurrence of `e`
86 self.map
[e
] = self[e
] + 1
90 # Count one more for each element of `es`
91 fun inc_all
(es
: Collection[E
])
96 # A new Counter initialized with `inc_all`.
97 init from
(es
: Collection[E
])
102 # Return an array of elements sorted by occurrences
105 # var c = new Counter[String].from(["a", "a", "b", "b", "b", "c"])
106 # assert c.sort == ["c", "a", "b"]
110 var res
= map
.keys
.to_a
111 var sorter
= new CounterComparator[E
](self)
116 # The method used to display an element
117 # @toimplement by default just call `to_s` on the element
118 protected fun element_to_s
(e
: E
): String
123 # Display statistical information
127 print
" population: {list.length}"
128 if list
.is_empty
then return
129 print
" minimum value: {self[list.first]}"
130 print
" maximum value: {self[list.last]}"
131 print
" total value: {self.sum}"
132 print
" average value: {div(self.sum,list.length)}"
133 print
" distribution:"
136 var limit
= self[list
.first
]
138 if self[t
] > limit
then
139 print
" <={limit}: sub-population={count} ({div(count*100,list.length)}%); cumulated value={sum} ({div(sum*100,self.sum)}%)"
142 while self[t
] > limit
do
144 if limit
== 0 then limit
= 1
150 print
" <={limit}: sub-population={count} ({div(count*100,list.length)}%); cumulated value={sum} ({div(sum*100,self.sum)}%)"
153 # Display up to `count` most used elements and `count` least used elements
154 # Use `element_to_s` to display the element
155 fun print_elements
(count
: Int)
160 if list
.length
<= count
*2 then min
= list
.length
162 var t
= list
[list
.length-i-1
]
163 print
" {element_to_s(t)}: {self[t]} ({div(self[t]*100,self.sum)}%)"
165 if list
.length
<= count
*2 then return
168 var t
= list
[min-i-1
]
169 print
" {element_to_s(t)}: {self[t]} ({div(self[t]*100,self.sum)}%)"
173 # Return the element with the highest value (aka. the mode)
176 # var c = new Counter[String].from(["a", "a", "b", "b", "b", "c"])
177 # assert c.max == "b"
180 # If more than one max exists, the first one is returned.
181 fun max
: nullable E
do
182 var max
: nullable Int = null
183 var elem
: nullable E
= null
185 if max
== null or v
> max
then
193 # Return the couple with the lowest value
196 # var c = new Counter[String].from(["a", "a", "b", "b", "b", "c"])
197 # assert c.min == "c"
200 # If more than one min exists, the first one is returned.
201 fun min
: nullable E
do
202 var min
: nullable Int = null
203 var elem
: nullable E
= null
205 if min
== null or v
< min
then
213 # Values average (aka. arithmetic mean)
216 # var c = new Counter[String].from(["a", "a", "b", "b", "b", "c"])
217 # assert c.avg == 2.0
220 if values
.is_empty
then return 0.0
221 return (sum
/ values
.length
).to_f
224 # The standard derivation of the counter values
227 # var c = new Counter[String].from(["a", "a", "b", "b", "b", "c"])
228 # assert c.std_dev > 0.81
229 # assert c.std_dev < 0.82
231 fun std_dev
: Float do
234 for value
in map
.values
do
235 sum
+= (value
.to_f
- avg
).pow
(2.0)
237 return (sum
/ map
.length
.to_f
).sqrt
241 private class CounterComparator[E
: Object]
243 redef type COMPARED: E
244 var counter
: Counter[E
]
245 redef fun compare
(a
,b
) do return self.counter
.map
[a
] <=> self.counter
.map
[b
]
249 private fun show_counter
(c
: Counter[Int])
252 default_comparator
.sort
(list
)
254 print
" {e} -> {c[e]} times ({div(c[e]*100, c.sum)}%)"
258 # Display exhaustive metrics about the poset
261 var nb_greaters
= new Counter[E
]
262 var nb_direct_greaters
= new Counter[E
]
263 var nb_smallers
= new Counter[E
]
264 var nb_direct_smallers
= new Counter[E
]
265 var nb_direct_edges
= 0
269 nb_edges
+= ne
.greaters
.length
270 nb_direct_edges
+= ne
.direct_greaters
.length
271 nb_greaters
[n
] = ne
.greaters
.length
272 nb_direct_greaters
[n
] = ne
.direct_greaters
.length
273 nb_smallers
[n
] = ne
.smallers
.length
274 nb_direct_smallers
[n
] = ne
.direct_smallers
.length
276 print
"Number of nodes: {self.length}"
277 print
"Number of edges: {nb_edges} ({div(nb_edges,self.length)} per node)"
278 print
"Number of direct edges: {nb_direct_edges} ({div(nb_direct_edges,self.length)} per node)"
279 print
"Distribution of greaters"
280 nb_greaters
.print_summary
281 print
"Distribution of direct greaters"
282 nb_direct_greaters
.print_summary
283 print
"Distribution of smallers"
284 nb_smallers
.print_summary
285 print
"Distribution of direct smallers"
286 nb_direct_smallers
.print_summary
290 # Helper function to display `n/d` and handle division by 0
291 fun div
(n
: Int, d
: Int): String
293 if d
== 0 then return "na"
294 return ((100*n
/d
).to_f
/100.0).to_precision
(2)